Index: lib/Transforms/Instrumentation/MemorySanitizer.cpp
===================================================================
--- lib/Transforms/Instrumentation/MemorySanitizer.cpp
+++ lib/Transforms/Instrumentation/MemorySanitizer.cpp
@@ -3249,8 +3249,10 @@
   // An unfortunate workaround for asymmetric lowering of va_arg stuff.
   // See a comment in visitCallSite for more details.
   static const unsigned AMD64GpEndOffset = 48;  // AMD64 ABI Draft 0.99.6 p3.5.7
-  static const unsigned AMD64FpEndOffset = 176;
+  static const unsigned AMD64FpEndOffsetSSE = 176;
+  static const unsigned AMD64FpEndOffsetNoSSE = 48;
 
+  unsigned AMD64FpEndOffset;
   Function &F;
   MemorySanitizer &MS;
   MemorySanitizerVisitor &MSV;
@@ -3262,7 +3264,18 @@
   enum ArgKind { AK_GeneralPurpose, AK_FloatingPoint, AK_Memory };
 
   VarArgAMD64Helper(Function &F, MemorySanitizer &MS,
-                    MemorySanitizerVisitor &MSV) : F(F), MS(MS), MSV(MSV) {}
+                    MemorySanitizerVisitor &MSV)
+      : F(F), MS(MS), MSV(MSV) {
+    AMD64FpEndOffset = AMD64FpEndOffsetSSE;
+    for (const auto &Attr : F.getAttributes().getFnAttributes()) {
+      if (Attr.isStringAttribute() &&
+          (Attr.getKindAsString() == "target-features")) {
+        if (Attr.getValueAsString().contains("-sse"))
+          AMD64FpEndOffset = AMD64FpEndOffsetNoSSE;
+        break;
+      }
+    }
+  }
 
   ArgKind classifyArgument(Value* arg) {
     // A very rough approximation of X86_64 argument classification rules.
Index: test/Instrumentation/MemorySanitizer/msan_basic.ll
===================================================================
--- test/Instrumentation/MemorySanitizer/msan_basic.ll
+++ test/Instrumentation/MemorySanitizer/msan_basic.ll
@@ -915,6 +915,26 @@
 ; CHECK: call void (i32, ...) @VAArgStructFn
 ; CHECK: ret void
 
+; Same code compiled without SSE (see attributes below).
+; The register save area is only 48 bytes instead of 176.
+define void @VAArgStructNoSSE(%struct.StructByVal* nocapture %s) sanitize_memory #0 {
+entry:
+  %agg.tmp2 = alloca %struct.StructByVal, align 8
+  %0 = bitcast %struct.StructByVal* %s to i8*
+  %agg.tmp.sroa.0.0..sroa_cast = bitcast %struct.StructByVal* %s to i64*
+  %agg.tmp.sroa.0.0.copyload = load i64, i64* %agg.tmp.sroa.0.0..sroa_cast, align 4
+  %agg.tmp.sroa.2.0..sroa_idx = getelementptr inbounds %struct.StructByVal, %struct.StructByVal* %s, i64 0, i32 2
+  %agg.tmp.sroa.2.0..sroa_cast = bitcast i32* %agg.tmp.sroa.2.0..sroa_idx to i64*
+  %agg.tmp.sroa.2.0.copyload = load i64, i64* %agg.tmp.sroa.2.0..sroa_cast, align 4
+  %1 = bitcast %struct.StructByVal* %agg.tmp2 to i8*
+  call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 4 %1, i8* align 4 %0, i64 16, i1 false)
+  call void (i32, ...) @VAArgStructFn(i32 undef, i64 %agg.tmp.sroa.0.0.copyload, i64 %agg.tmp.sroa.2.0.copyload, i64 %agg.tmp.sroa.0.0.copyload, i64 %agg.tmp.sroa.2.0.copyload, %struct.StructByVal* byval align 8 %agg.tmp2)
+  ret void
+}
+
+attributes #0 = { "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+fxsr,+x87,-aes,-avx,-avx2,-avx512bitalg,-avx512bw,-avx512cd,-avx512dq,-avx512er,-avx512f,-avx512ifma,-avx512pf,-avx512vbmi,-avx512vbmi2,-avx512vl,-avx512vnni,-avx512vpopcntdq,-f16c,-fma,-fma4,-gfni,-pclmul,-sha,-sse,-sse2,-sse3,-sse4.1,-sse4.2,-sse4a,-ssse3,-vaes,-vpclmulqdq,-xop,-xsave,-xsaveopt" "unsafe-fp-math"="false" "use-soft-float"="false" }
+
+; CHECK: bitcast { i32, i32, i32, i32 }* {{.*}}@__msan_va_arg_tls {{.*}}, i64 48
 
 declare i32 @InnerTailCall(i32 %a)